Muscle Anatomy-aware Geometric Deep Learning for sEMG-based Gesture Decoding

2025-10-20

AI Summary

To address the challenges of modeling multidimensional non-stationary electromyographic (sEMG) time series and poor cross-subject/cross-session generalization in gesture decoding, this paper proposes a geometric deep learning framework grounded on the Symmetric Positive Definite (SPD) manifold. Methodologically, we design an SPD manifold embedding module to capture non-Euclidean representations of muscle activation dynamics; introduce domain-specific batch normalization coupled with unsupervised domain adaptation to mitigate inter-subject and inter-session variability without re-calibration; and integrate multi-kernel temporal feature extraction to enhance discriminability. Evaluated on Ninapro DB6 and Flexwear-HD, our method achieves cross-session classification accuracy improvements of 8.83% and 4.63%, respectively—outperforming existing Euclidean and mainstream manifold-based approaches and establishing new state-of-the-art performance. These results validate the efficacy of jointly leveraging manifold geometric priors and domain adaptation for robust sEMG gesture decoding.

Technology Category

• Machine Learning: Learning with Manifolds
• Computer Vision: Biometrics, Face, Gesture & Pose
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

Robust and accurate decoding of gesture from non-invasive surface electromyography (sEMG) is important for various applications including spatial computing, healthcare, and entertainment, and has been actively pursued by researchers and industry. Majority of sEMG-based gesture decoding algorithms employ deep neural networks that are designed for Euclidean data, and may not be suitable for analyzing multi-dimensional, non-stationary time-series with long-range dependencies such as sEMG. State-of-the-art sEMG-based decoding methods also demonstrate high variability across subjects and sessions, requiring re-calibration and adaptive fine-tuning to boost performance. To address these shortcomings, this work proposes a geometric deep learning model that learns on symmetric positive definite (SPD) manifolds and leverages unsupervised domain adaptation to desensitize the model to subjects and sessions. The model captures the features in time and across sensors with multiple kernels, projects the features onto SPD manifold, learns on manifolds and projects back to Euclidean space for classification. It uses a domain-specific batch normalization layer to address variability between sessions, alleviating the need for re-calibration or fine-tuning. Experiments with publicly available benchmark gesture decoding datasets (Ninapro DB6, Flexwear-HD) demonstrate the superior generalizability of the model compared to Euclidean and other SPD-based models in the inter-session scenario, with up to 8.83 and 4.63 points improvement in accuracy, respectively. Detailed analyses reveal that the model extracts muscle-specific information for different tasks and ablation studies highlight the importance of modules introduced in the work. The proposed method pushes the state-of-the-art in sEMG-based gesture recognition and opens new research avenues for manifold-based learning for muscle signals.

Problem

• Decoding gestures from non-invasive sEMG signals robustly and accurately
• Addressing high variability across subjects and sessions in sEMG decoding
• Analyzing multi-dimensional non-stationary sEMG time-series with long-range dependencies

Innovation

• Geometric deep learning on SPD manifolds for sEMG
• Unsupervised domain adaptation for cross-session robustness
• Multi-kernel feature projection with batch normalization